Statistics Glossary → Number needed to treat (NNT)
Number needed to treat (NNT)concept
The number of people who must receive a treatment, compared with the control condition, to produce one additional good outcome (or prevent one additional bad event).
The NNT is the reciprocal of the absolute risk difference: NNT = 1 / |p_control - p_treatment|. If the event rate is 20% in the control group and 15% in the treatment group, for example, the absolute risk reduction is 0.05 and NNT = 1 / 0.05 = 20; on average, treating 20 people prevents one event. The value is usually rounded up to the next whole number. When the treatment causes harm, the same calculation is reported as the 'number needed to harm' (NNH).
The NNT is clinically easier to grasp than the relative risk or odds ratio because it shows the absolute size of the effect: the same relative risk reduction gives a small NNT in a high-risk group and a large one in a low-risk group. An NNT therefore always refers to a specific follow-up period and baseline risk and should not be compared directly across studies. If the confidence interval for the risk difference includes zero, the confidence interval for the NNT passes through infinity and must be interpreted with care.
An everyday example
Suppose a vaccine lowers the chance of catching flu over one winter from 10% to 4%. The absolute risk difference is 0.06, so NNT = 1 / 0.06 ≈ 16.7, or about 17: roughly 17 people must be vaccinated to prevent one case of flu. In a community where the flu rate is only 1% to begin with, the same relative effect would give a much larger NNT.
In MerQur
It is offered in the Effect Size Calculator among the effect sizes computed from the event rates of two groups.